SkillFoxx

Agent Autoloop CLI

CLI dispatcher for agents: tasks run in the background and the agent sleeps on a question until you reply from a messenger.

  • 4,572 ★ on the repository
  • Last commit
  • MIT
CLI

Broad access

Why this level

  • The agent works unattended and the installer runs as an internet script.
All reasons and checks

kkkkhazix/agent-autoloopAuthor:Khazix

Install

Manual install

irm https://raw.githubusercontent.com/KKKKhazix/agent-autoloop/main/install.ps1 | iex

The first README option, for Windows PowerShell; macOS and Linux get a `curl -fsSL .../install.sh | sh` command.

Access and checks

How we review

Broad accessThe tool can change data, run commands or handle secrets.

  • The agent works unattended and the installer runs as an internet script.

This is third-party code. Review the repository files before installing.

What it does

Agent Autoloop is a small automation system for command-line agents such as Claude Code and Codex. A dependency-free kernel schedules tasks, while event sources, agents, notification channels and even code shipping plug in as plugins. The agent works unattended and, when unsure, sends a card with options to Feishu, DingTalk, WeCom, Slack, Discord or a desktop notification. Until you answer, the agent holds no process and spends no tokens. After the reply from `agent-autoloop answer` it resumes the same session. `agent-autoloop status` shows overall progress. A new agent plugs in with one config line: any tool that takes a prompt and prints a final reply works. The repository has scenario templates and plugin examples.

Who it is for. Developers and small teams who want an agent to work without constant watching and to ask a human only about hard calls.

Good fit when

  • An agent processes tracker tasks and ambiguous calls need answers in a messenger.
  • Work runs on a laptop and a cloud server is not wanted.
  • A second agent with only a console call must be plugged in.
  • A one-command status summary is needed in the morning.

Not a fit when

  • You need a web panel with task lists and access roles.
  • The node must be reachable from outside as a service.
  • Decisions cannot be left to an agent even while waiting for an answer.

Example request

Run `agent-autoloop try`, answer the sample colour question and check with `agent-autoloop status` that the agent finished.

Limitations

It needs Node 24 or newer, Git and an already signed-in command-line agent; macOS and Linux also need curl. The installer runs as an internet script through `curl ... | sh` or `irm ... | iex`, so open and read it first. There is no browser UI: status comes from a command and decisions arrive as messenger cards. Feishu, DingTalk and WeCom channels need those services keys. Documentation is mostly Chinese, with a short English README.

How to disable. Stop the service, remove the package and the settings directory. Follow the project docs for removal and revoke channel keys in the services themselves.

FAQ

How do I install it?

On PowerShell: `irm https://raw.githubusercontent.com/KKKKhazix/agent-autoloop/main/install.ps1 | iex`. For macOS and Linux the README gives a `curl ... | sh` command. An agent can also be handed `INSTALL.md`.

How do I verify it?

Run `agent-autoloop try`: it starts a trial task, asks a colour question and the reply is given with `agent-autoloop answer`.

Which agents fit?

Any command-line tool that takes a prompt and prints the last reply. Claude Code and Codex are the main examples in the README.

Is an open port needed?

No. It runs on one computer and needs no public port or cloud server.

How do I add my own agent?

With one config line and no SDK. Ready samples of agents, channels and event sources sit in `examples/plugins/`.

Discussion

No questions or reviews yet.

CLI tools
Official

A self-hosted control plane for running several agents: task dispatch, run review, spend tracking, connected via MCP, CLI or REST

MCP serverBroad access6.5K

Connects Claude Code, Cursor, Codex or OpenCode to Telegram, Slack and Discord: teammates message a bot, an agent replies

CLIBroad access325

Agents for recurring work in Markdown files: schedules, approval of important actions, run history in a dashboard and a Mac app

CLIExtended access208

An agent messages your phone or calls it over VoIP, a voice AI discusses the decision and the transcript returns to the agent

SkillBroad access401

A phone API for AI agents: Claude and Cursor get a real number, make and take calls, read SMS and transcripts over MCP.

MCP serverBroad access62

A plugin and CLI catalog for agents: generates command-line interfaces for GUI apps like GIMP and Blender and installs ready ones via CLI-Hub

PluginBroad access58.6K

README in short

The README lays out the problem, quick install, a trial run, scenario templates, plugins, kernel mechanisms and links to docs.

README badge

Report a problem